Disjoint regions sudoku [new example]

[Fred76]

Rules: Place a number from 1-9 in each empty cell in the grid such that each row, column and set of cells of same colour contains each number exactly once.

disjoint.png (15.07 KiB) Viewed 13607 times

It's a kind of disjoint groups sudoku without 3*3 boxes. It can be seen in another way, too; I'll comment more in few days, when some people have had time to solve it and shared their feelings.

[detuned]

Could this variant also be described by the rules of Scattered Sudoku?

[Fred76]

Tom, is it possible to make a poll on this one, too?
Thanks,
Fred

[Realshaggy]

I would refuse to solve it in any contest. This is just a lazy way of adding artificial difficulty to the puzzle and annoying to solve. Even more than using letters instead of numbers without special reasons to do so. Imho things like that are disrespectful to the solver.

[Fred76]

I would refuse to solve it in any contest. This is just a lazy way of adding artificial difficulty to the puzzle and annoying to solve. Even more than using letters instead of numbers without special reasons to do so. Imho things like that are disrespectful to the solver.

Just for curiosity, would you say that this puzzle is not a sudoku or that it is a sudoku but far to be appropriate for competition?

Fred

[puzzlemad]

Without getting into discussions about what is/isn't appropriate, my concern with this is the use of the colours - which could be a problem for those with colour blindness, unless of course the disjointed regions were marked some other way.

[Realshaggy]

It is clearly a sudoku but not appropriate for a contest in my opinion. And not only for the colour-problems.

[Feadoor]

This variant (with these exact regions) is pointless, because it's completely equivalent to classic sudoku. Take the yellow cells as the top-left 3x3 box of a classic, the red cells as the top-middle, the green cells as the top-right, the pink cells as the middle-left etc. Then all cells which were previously in the same row or column as each other still are, so the constraints haven't changed.

It transforms to the following classic:

+---------+---------+---------+
| .  .  . | .  .  . | .  2  4 |
| .  .  . | .  7  8 | .  .  1 |
| .  .  . | .  .  6 | 5  .  . |
+---------+---------+---------+
| .  .  . | .  .  . | 6  7  . |
| .  5  . | .  .  . | .  8  . |
| .  4  3 | .  .  . | .  .  . |
+---------+---------+---------+
| .  .  2 | 3  .  . | .  .  . |
| 1  .  . | 4  5  . | .  .  . |
| 6  7  . | .  .  . | .  .  . |
+---------+---------+---------+

I for one would probably be faster at this puzzle by first drawing out this classic and solving that instead.

[detuned]

So I think this comes down to what we are discussing in another thread. We clearly have rows, columns and regions, but I think it's debatable whether in the original presentation the regions are clearly marked.

Liane's point about the use of colour is interesting - yes we have a colour blindness issue going on here, but we also have a printing issue given that almost every printed competition I can think of has been done in greyscale.

This clearly will not work for this puzzle. Disjoint groups is typically presented with no shading when printed - if you were to do the same for this puzzle then it's very hard to argue the regions had been clearly marked.

[Feadoor]

I think, even if this puzzle could be presented in a way that made the regions clear, I would argue against its inclusion on the grounds that it is entirely equivalent to classic sudoku and this presentation of classic sudoku amounts to nothing more than a gimmick.

[Fred76]

I basically agree with everything that has been said here.
Yes the use of colours is a problem. We can discuss about clarity of marking regions (I feel it's hard to make it clear without use of colours, but perhaps not impossible).
Feadoor exactly understood what I did, and reconstructed the classic sudoku which was the reference for this one: http://sudokuvariante.blogspot.com/2016 ... e-n88.html Congrats !
I'm sorry to have created a puzzle which was not fun to solve, but I think it was useful for the debate here.

I've another arguments against this puzzle: I feel regions should be something well localized. In other words constituted by connected cells.
To be fair, I've no problem with scattered sudoku, which has one region which is scattered in the whole grid, but still has 8 regions made by connected cells. The note concern only primary regions, I think if there are located primary regions, we can add extra-regions which are scattered, my point is not to say disjoint group sudoku is not appropriate.

Fred

[TiiT]

I think it's a latin square with extra regions.